What problem does it solve? It prevents premature "done" declarations by requiring verifiable evidence—executed commands, exit codes, timestamps, and scope—before any work can be reported as complete. ## Core Features & Use Cases - Evidence-Based Completion: Requires recorded commands, exit codes, execution times, and target scope instead of verbal claims of success. - Staged Verification Pipeline: Enforces a four-stage sequence from focused tests through full test suites, static analysis, and end-to-end integration checks. - Mode-Specific Definitions of Done: Applies different completion checklists for light, operation, standard, legacy-refactor, and documentation change modes. - Use Case: Before reporting a bug fix as complete, run the focused test, the full test suite, and the linter, then record each command with its exit code in the verification report table. ## Quick Start Ask the AI to verify whether the current change meets the completion criteria and produce a verification evidence report with commands and exit codes.
